Entamoeba histolytica infection is extremely common and affects an estimated 480 million individuals annually. However, only about 10% of these persons develop symptoms such as colitis or liver abscess. The low incidence of symptom occurrence is putatively due to the existence of both pathogenic and nonpathogenic forms of the amoeba. As of 1988, it had been established that the subjects who eventually exhibit symptoms harbor "zymodemes" which have been classified as such on the basis of their distinctive hexokinase and phosphoglucomutase isoenzymes. The pathogenic forms are not, however, conveniently distinguishable from the nonpathogenic counterparts using morphogenic criteria.
The distinction between pathogenic and nonpathogenic strains in diagnosis is of great practical importance, because only persons infected with E. histolytica who will develop the disease should be treated. This is bad enough in developed countries where it would at least be possible economically to treat every carrier with a known effective drug (metronidazole); it is, of course, undesirable to administer such drugs unnecessarily. In less developed countries, the cost of these unnecessary administrations is significant enough to have a dramatic negative impact on the resources for overall health care.
There is an almost perfect correlation between infection with a pathogenic zymodeme and development of symptoms and between infection with a nonpathogenic zymodeme and failure to develop these symptoms. As a general proposition, only pathogenic strains can be grown axenically (i.e., in the absence of an associated microorganism) and nonpathogenic strains have been made to grow in this manner only by "training" them to do so in a series of media alterations beginning with attenuated bacteria. The adaptation was accompanied by exhibition of the enzyme pattern characteristic of pathogenic strains (Mirelman, D., et al., Infect Immun (1986) 54:827-832). This work has not been repeatable in other laboratories, and more recent work on genomic differences (see below) indicates that the pathogenic and nonpathogenic forms are separate species.
It is known that E. histolytica infection is mediated at least in part by the "Gal/GalNAc" adherence lectin which was isolated from a pathogenic strain and purified 500 fold by Petri, W. A., et al., J Biol Chem (1989) 264:3007-3012. This successful isolation and purification was preceded by the production of mouse monoclonal antibodies which inhibit the in vitro adherence of the amoebic trophozoites; the antibodies were prepared from immortalized cells from spleens of mice immunized with sonicated trophozoites grown in axenic culture after having originally been isolated as a pathogenic strain from an affected subject. (Ravdin, J. I., et al., Infect Immun (1986) 53:1-5.) The cells were screened by the ability of the supernatants to inhibit adherence of the trophozoites to target tissue. All of these reported monoclonal antibodies, therefore, are presumably immunoreactive with the Gal/GalNAc surface adhesion of the pathogen. The Gal/GalNAc lectin was then prepared by galactose affinity chromatography and reported in 1987. (Petri, W. J., et al., J Clin Invest (1987) 80:1238-1244). Studies of serological cross-reactivity among patients having symptomology characteristic of E. histolytica pathogenic infection, including liver abscess and colitis, showed that the adherence lectin was recognized by all patients' sera tested (Petri, Jr., W. A., et al., Am J Med Sci (1989) 296:163-165).
The purified "Gal/GalNAc" lectin was shown to have a nonreduced molecular weight of 260 kd on SDS-PAGE. After reduction, with beta-mercaptoethanol, the lectin separated into two subunits of 170 and 35 kd MW. Further studies showed that antibodies directed to the 170 kd subunit were capable of blocking surface adhesion to test cells, and therefore, the 170 kd subunit is believed to be of primary importance in mediating adhesion.
Despite the generally interesting and useful results cited above, the ability to diagnose the presence or absence of pathogenic strains of E. histolytica has proved difficult. Since both pathogens and nonpathogens are morphologically similar, microscopic tests are not particularly useful. ELISA techniques have been used to detect the presence or absence of E. histolytica antigen in both stool specimens and in sera, but these tests do not seem to distinguish between the pathogenic and nonpathogenic strains. Root et al., Arch Invest Med (Mex) (1978) 9: Supplement 1:203, pioneered the use of ELISA techniques for the detection of amoebic antigen in stool specimens using rabbit polyclonal antiserum. Various forms of this procedure have been used since, some in correlation with microscopic studies, and all using polyclonal antisera. None of these, apparently, pinpoints the instances of infection with the pathogenic as opposed to nonpathogenic form. See, for example, Palacios et al., Arch Invest Med (Mex) (1978) 9: Supplement 1:203; Randall et al., Trans Roy Soc Trop Med Hyg (1984) 78:593; Grundy, Trans Roy Soc Trop Med Hyg (1982) 76:396; Ungar, Am J Trop Med Hyg (1985) 34:465.
These studies on stool specimens are summarized in Amebiasis: Human Infections by Entamoeba Histolytica, J. Ravdin, ed. (1988) Wiley Medical Publishing, pp. 646-648. Similar methods to detect characteristic E. histolytica antigens in serum and in liver abscess fluid are equally unable to distinguish pathogens from non-pathogens (ibid., pp. 661-663). As summarized in this article, as of 1988, the only known way to distinguish pathogenic from nonpathogenic forms of this amoeba was through characterizing the isoenzyme pattern using electrophoresis.
Recently it has been shown by two different groups that differences between pathogenic and nonpathogenic strains can be demonstrated using comparisons of DNA isolates. Garfinkel, L. I., et al., Infect Immun (1989) 57:926-931 developed DNA probes which hybridize to DNA isolated from E. histolytica and four types of restriction fragment length patterns were obtained. These patterns correlated with pathogenic/nonpathogenic distinctions. Similarly, Tannich, E., et al., Proc Natl Acad Sci (1989) 86:5118-5122 probed cDNA libraries constructed from various strains and showed that pathogenic isolates were genetically distinct from nonpathogenic ones. However, these techniques require the culture of the organisms isolated from patients to obtain sufficient quantities for testing, and are thus time consuming and labor intensive.
Strachan, W. D., et al., Lancet (1988) 561-562, report the production of two monoclonal antibodies designated 22.3 and 22.5 which were members of a large group prepared by standard procedures from mice immunized with axenic cultures of a pathogenic E. histolytica strain NIH200/ATCC 30458. These monoclonal antibodies were tested in an immunofluorescence assay with cultures obtained from both putatively invasive and noninvasive strains, and appeared to immunoreact only with culture samples of invasive strains. There is no indication in this publication as to the manner of screening for antibodies with this characteristic, it is not known to what target these antibodies bond, nor would it be possible, without these specific antibodies, to reproduce this result. The test described requires intact E. histolytica and therefore cannot be applied in serum, urine or liver abscess fluid and can only be applied to stool samples which are freshly collected.
It would be highly desirable to have a relatively simple clinical test which would detect the presence of E. histolytica antigen in samples taken from patients and to be able to distinguish whether these antigens are associated with pathogenic or nonpathogenic strains. The present invention offers such a test, by employing conventional immunoassay procedures using monoclonal antibody reagents which are specifically immunoreactive with pathogenic and/or nonpathogenic Gal/GalNAc adherence lectin.